Global performance of metaheuristic optimization tools for water distribution networks

Modified Jaya Algorithm for Optimal Design of Water Distribution Network. (Dept. C)

“…Any algorithm performance depends on the following points: 1) Effectiveness: the proximity of the optimal solution to the global known solution, 2) Efficiency: the number of objective function evaluations necessary to reach the optimum solution and the total permissible objective function evaluations in any run, and 3) Reliability: the ratio of the number of runs reached the global solution to the total number of runs. Recently, Djebedjian et al [16] suggested two criteria to investigate the performance of different metaheuristic algorithms in case of using one or multiple runs. Then, they apply their criteria to compare between different algorithms using all the available WDN's literature data.…”

“…Three benchmarks WDN's in addition to one national WDN are solved in the present work to investigate the performance of the proposed FSAJA. The three benchmark WDN's, used by Djebedjian et al, [16], are adopted here for the sake of comparing the performance of FSAJA to performances of other algorithms presented in [16]. The studied WDN's are: 1) the two-loop WDN, 2) the New York WDN, and 3) the Hanoi WDN.…”

“…(7); 2) the image strategy is adopted to replace the generated violated decision variable with an interior one within its permissible limits, Eqs. (16) or (17); 3) population size is fixed and equal to four times number of the decision variables; 4) the penalty is assumed initially very big value and it is self-adjusted with the generation process, Eq. (20), to a lower magnitude that gives infeasible candidates around the solution domain, the opportunity of sharing their information's in the next generations; 5) fittest feasible candidate in a generation cannot be replaced, with a generated more fittest infeasible one, and 6) satisfying any of the following two stopping criteria to terminate the run: a) if the ratio between the standard deviation of candidates objective functions and their average, goes to be less than 0.0001 at any generation, or b) the processed number of generations without improvement in the fittest candidate is more than 30, Eq.…”

A Metaheuristic Algorithm for Flexible Energy Storage Management in Residential Electricity Distribution Grids

“…Metaheuristic algorithms such as the Genetic Algorithm, Particle Swarm Optimization, Differential Evolution, Whale Algorithm, Fireworks Algorithm, are nature-inspired optimization techniques that proved their performance in problems regarding the optimization of operation conditions in electrical networks. In the latest years, they were applied for integrating electric vehicles and distributed generation into smart grids [36], optimal reconfiguration of distribution networks [37], optimal power flow analysis in DC distribution networks [38], reliability improvement [39], optimal consumption planning [40]. In the paper, a Genetic Algorithm was used to determine the optimal buses and phases of connection for a fixed number of storage units (batteries), with the aim of reducing the energy losses over a time interval of 24 hours, according to the assumptions and scenarios described in Section 1.…”

Cost Optimization of Water Distribution Networks without Storage Tanks

“…This is achieved by minimizing the operating and capital costs of these components while the constraints (minimum nodal pressure and demand) at the consumer nodes are fulfilled and the hydraulic laws (mass and energy conservations) are maintained. General recent and thorough reviews on the water distribution network (WDN) optimization can be found [1][2][3][4][5][6][7][8]. …”